Provides natural-language statistical analysis via MCP, enabling regression, time series, ML, and data visualization through an interactive Python-based server.
Configuration
View docs{
"mcpServers": {
"gojiplus-rmcp": {
"url": "https://rmcp-server-394229601724.us-central1.run.app/mcp",
"headers": {
"RMCP_HTTP_PORT": "9000",
"RMCP_LOG_LEVEL": "DEBUG",
"RMCP_R_TIMEOUT": "180"
}
}
}
}RMCP enables you to perform comprehensive statistical analysis through natural-language conversations. You interact with an MCP server from your preferred client, ask for regression, time-series, machine learning, or data-wrangling tasks in plain language, and receive structured results, including markdown tables and inline visuals.
Interact with the RMCP server using your MCP client (Claude Desktop, Claude Web, or any compatible client). Start a session, describe the analysis you want in plain language, and review the results that include model summaries, statistics, and inline visualizations.
Practical usage patterns you can try include: - Describe a regression or forecast task and let RMCP fit a model, report coefficients, p-values, and goodness-of-fit measures. - Request time-series analyses such as ARIMA, decomposition, or forecasting for a specified horizon. - Ask for machine-learning-oriented tasks like clustering or classification to segment or predict outcomes. - Ask for descriptive statistics, correlations, or outlier checks to understand your data before modeling.
Prerequisites you need before installing RMCP: - Python 3.10 or newer - R 4.4.0 or newer (for the full R package ecosystem used by RMCP)
Install RMCP from PyPI and enable HTTP transport support if you plan to use web clients:
pip install rmcp
pip install rmcp[http]If you want to develop RMCP locally from source, clone the repository and install in editable mode:
git clone https://github.com/finite-sample/rmcp.git
cd rmcp
pip install -e ".[dev]"Configuration options, security considerations, examples, and troubleshooting are available to tailor RMCP to your environment. You can configure HTTP or stdio transports, run in development or production modes, and set environment variables to control ports, timeouts, and log levels.
Configuration and transport options include HTTP server endpoints for web clients and stdio configurations for Claude Desktop. You can also use a local HTTP server for testing and development.
Two example connection methods are shown below. Use the HTTP URL to connect web clients and the stdio command to run RMCP locally from your development environment.
Troubleshooting tips cover common issues like missing R packages, MCP connection problems, and verifying installation with a version check.
Perform linear and generalized linear regression with summaries, coefficients, p-values, R-squared, and diagnostic statistics.
Fit ARIMA, decomposition, and related models to forecast future values and assess accuracy.
Train clustering, decision trees, random forests, and related models with built-in evaluation metrics.
Run T-tests, ANOVA, chi-square tests, and normality assessments to support hypothesis testing.
Compute descriptive statistics, correlations, and data summaries for exploratory analysis.
Apply standardization, winsorization, and lag/lead feature construction for modeling prep.
Generate inline plots and visualizations such as scatter plots, histograms, and heatmaps for Claude outputs.
Import CSV, Excel, and JSON data with validation checks and schema interpretation.
Assist in constructing statistical formulas and model specifications from natural language.
Handle input errors gracefully with suggested corrections and fallbacks.